Hello,

I understand that the correlation length (in a particular dimension) is the distance (in that dimension) over which an observation influences its neighborhood.

In the documentation, there is the https://gher-uliege.github.io/DIVAnd.jl/latest/index.html#DIVAnd.DIVAnd_kernel function with the following comment:

K(r) is the kernel function (function of the normalized distance r), len_scale is the distance at which K(len_scale) = 0.6019072301972346 (which is besselk(1,1))

I have the following questions:

• Is there a connection between len_scale and the correlation length?
• Considering that the kernel value is 1 at a particular observation, what is the value of the kernel at the correlation length?
• Knowing this value, for instance 0.5, and considering a particular case where there is only 1 observation equal to X, with a background equal to 0, does it mean that the analysis will give a value of 0.5 * X at the correlation length if epsilon2 is close to 0?
• I understand that a Bessel kernel is used, am I correct?

Maybe such information, especially the value of the kernel at the correlation length, would be interesting in the documentation in order to have an accurate idea of the magnitude of the influence of the observation at the correlation length.

Thank you in advance for your help.